TWI656340B - Automated decision-based energy-dispersive x-ray methodology and apparatus - Google Patents

Abstract

One embodiment relates to a method for automatically redetecting defects detected in a defective die on a target substrate. The method comprises: performing a primary re-detection of one of the defects using a primary electron microscope (SEM) to obtain an electron beam image of the defects; performing a form based on the defects as determined from the electron beam images The defects are automatically classified into one of several types; a particular type of defect is selected for automatic energy dispersive x-ray (EDX) re-detection; and the automatic EDX re-detection is performed for the particular type of such defect. In addition, an automated technique for obtaining an accurate reference to improve the usefulness of EDX results is disclosed. In addition, an automatic method for classifying the defects based on the EDX results is disclosed, the automatic method providing one of the combined form information and element information, and finally a pareto. Other embodiments, aspects, and features are also disclosed.

Description

Method and device for automatically determining energy-based X-rays Cross-reference to related applications

The present patent application claims the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the disclosure. This patent application also claims the benefit of U.S. Provisional Patent Application Serial No. Ser. This patent application also claims the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the disclosure. The present patent application also claims the benefit of US Provisional Patent Application Serial No. Serial No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. The present patent application claims priority to the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of the benefit of

The present invention relates to methods and apparatus for automated inspection and re-detection of substrates such as semiconductor wafers using energy dispersive x-ray spectroscopy.

Used in a scanning electron microscope (SEM)-based inspection instrument One of the secondary electron emission of the surface of the substrate focuses the electron beam to scan a fabricated substrate (such as a germanium wafer or a scale reticle). The emitted electrons are detected and the detected data is typically converted into an image of the surface of the sample. These images are then numerically analyzed to detect anomalies (referred to as defects) in the fabricated substrate. The detected defects can then be detected by further imaging.

The detected defects may also be manually or automatically classified into different categories or categories. The classification of a defect can be used to determine its cause so that appropriate adjustments can be made in the manufacturing process to improve its yield.

In addition to generating secondary electrons, an electron beam striking one of the samples in an SEM also produces x-rays characteristic of the material of the sample. In energy dispersive x-ray (EDX) spectroscopy, a solid state detector is positioned relatively close to the sample to collect x-rays emitted from the sample due to collision of electron beams. The detector receives and detects x-rays of different energies to obtain an energy spectrum of one of the detected x-rays. This energy spectrum provides information about the composition of the elements of the material that is irradiated with electrons.

One embodiment relates to a method for automatically redetecting defects detected in a defective die on a target substrate. The method comprises: obtaining a result file of one of the locations including the defects; performing an automatic re-detection of one of the defects using a primary electron microscope (SEM) to obtain an electron beam image of the defects; performing the electron based on the electrons The image of the defects determined by the beam image and the defects are automatically classified into one of several types; a particular type of defect is selected for automatic energy dispersive x-ray (EDX) re-detection; and the particular type The automatic EDX retest is performed on the same defect.

Another embodiment is directed to an apparatus for automatically redetecting defects detected on a target substrate. The apparatus includes: an electron beam column for generating a primary electron beam and focusing the primary electron beam onto a surface of the target substrate; and a movable stage for holding the target substrate at the primary Under the electron beam; a deflector for making the Primary electron beam deflection; an electronic detector for detecting secondary electrons emitted from the surface of the target substrate due to collision of the primary electron beam; an x-ray detector configured to detect An x-ray emitted from the surface of the target substrate due to the collision of the primary electron beam; and a control system including a non-transitory data storage for storing computer readable code and data and further comprising One of the computer readable code processors. The computer readable program code includes instructions for: obtaining a result file comprising a location of the defects detected in a defective die on the target substrate; performing the defects Automated secondary electron microscopy (SEM) re-detection to obtain electron beam images of the defects; performing automatic classification of the defects into one of several types based on the morphology of the defects as determined from the electron beam images Selecting a particular type of defect for automatic energy dispersive x-ray (EDX) re-detection; and performing the automatic EDX re-detection for the particular type of such defect.

Another embodiment is directed to a method of automated energy dispersive x-ray (EDX) re-detection of defects on a defective die. The method is automated under the control of computer readable instructions and includes: moving to a defect site; obtaining an EDX spectrum from the defect site; moving from the defect site to a reference site; from the reference site Obtaining the EDX spectrum; and generating a difference spectrum from the EDX spectrum from the defect site and the EDX spectrum from the reference site.

In another embodiment, the method of performing automatic energy dispersive x-ray (EDX) re-detection of defects on a defective die is performed on defects indicated as being in a repeating cell array. In this case, the movement from the defective site to the reference site is performed by deflecting a primary electron beam in a direction by a cell size such that the reference site is in an adjacent cell corresponding to the defect bit. One of the locations.

In another embodiment, the method of performing automatic energy dispersive x-ray (EDX) re-detection of defects on a defective die is performed on defects indicated as being in a non-array patterned structure. In this case, the movement from the defect site to the execution is performed in the following manner The reference site: translating and holding a carrier of the target substrate to move a field of view of a scanning electron microscope from the defect site on the defective die to the reference site on the adjacent die.

Other embodiments, aspects, and features are also disclosed.

402‧‧‧ Concerned defects

404‧‧‧corresponding reference point

901‧‧‧ source

902‧‧‧Incoming electron beam/primary electron beam

904‧‧‧Viine filter

906‧‧‧ deflector

907‧‧ ‧focus lens

908‧‧‧ Objective / Final Lens

910‧‧‧ Target substrate/substrate sample

912‧‧‧Secondary electron beam/scattered electron beam

914‧‧‧Secondary electronic detection system

919‧‧‧Translatable stage

920‧‧‧ Energy Dispersive X-ray Detection System

950‧‧‧Instrument Control and Data Processing System/Control/Processing System

952‧‧‧Processor/Microprocessor/Microcontroller

954‧‧‧Data storage

955‧‧‧Computer readable code/computer readable instructions/code

956‧‧‧Information

957‧‧‧User interface

958‧‧‧Display system

1 is a flow chart of one of the EDX methods based on an automatic determination in accordance with an embodiment of the present invention.

2 is a flow diagram of one of the program flows for automatically subsampling a identified defect in accordance with an embodiment of the present invention.

3 is a flow diagram of one of the program flows for automatic EDX re-detection in an array mode in accordance with an embodiment of the present invention.

4 shows an electron micrograph of a defect of interest (DOI) and one of a neighboring cell in a neighboring cell, in accordance with an embodiment of the invention.

5 is a flow diagram of one of the program flows for automatic EDX re-detection in a non-array mode, in accordance with an embodiment of the present invention.

6 shows an electron micrograph of a defect of interest (DOI) and a corresponding reference point on a reference die, in accordance with an embodiment of the present invention.

7 is a flow diagram of one of an automated procedure for combining a morphological defect result with an element defect result to obtain a final Pareto, in accordance with an embodiment of the present invention.

Figure 8 is a diagram of one of the inventive procedures for comparing one of the existing procedures for automatic EDX re-detection and for automatic EDX re-detection, in accordance with an embodiment of the present invention.

9 is a schematic illustration of one of the scanning electron microscope devices having an energy dispersive x-ray (EDX) detection system in accordance with an embodiment of the present invention.

1 is a flow diagram of an EDX method 100 based on an automatic determination in accordance with an embodiment of the present invention. This method 100 can be performed after identifying the defect location by automatic verification. This automatic inspection can be performed by an identical or a separate SEM device, for example. The target substrate being inspected may be, for example, one of a plurality of dies formed thereon. Alternatively, the target substrate can be a proportional mask.

Referring to step 101, SEM automatic re-detection of the identified defects can be performed. The SEM automatic retest can involve high resolution SEM imaging of defects identified in a result file from one of the automated tests.

Referring to step 102, after the SEM is automatically re-detected, the form-based automatic classification can be performed. As depicted in the example bar chart on the right side of step 102, the defects can be classified into a number of defect types. The bar graph shows the defect frequency versus defect type. Three types (#1, #2, and #3) are shown, but the number of types can of course be any number. For example, a first defect type can be a hole, a second defect type can be a scratch, and a third defect type can be a particle. These are just a few examples of defect types. Of course, other and different types of defects can be utilized. For example, instead of having a particle defect type, there may be a large particle defect type and a small particle defect type depending on the size of the particle, or a circular particle defect type and a non-circular may exist depending on the shape of the particle. Shape type of particle defect.

Referring to step 104, one of the defects of one (or more) types for automatic EDX retesting is made. The selection of the type of defect for automatic EDX redetection can be pre-programmed or pre-configured to be automatically executed by executing the executable code of the EDX method 100 based on the automatic decision. As shown in the example bar chart on the right side of step 104, the selected type can be type #3. For example, type #3 may correspond to one type of defect associated with a particle, which may also be referred to as a "fall on" defect. This automatic selection of the type of defect for automatic EDX re-detection according to step 104 may be referred to herein as an automatic sub-sampling. Further explanation of automatic subsampling is provided below with respect to FIG.

After the automatic subsampling of step 104, an automatic EDX redetection 105 with an accurate reference can be performed on the selected type of defect. To obtain an accurate reference, automatic EDX re-detection 105 may involve repeating steps 106-116 for each defect of the selected type.

Referring to step 106, movement to a defect site can be performed. This movement to the defect site can be achieved by translating the stage such that the defect site is within the field of view of the SEM. This operation can be performed based on the position coordinates of the defect sites obtained from the stored results of the SEM automatic re-detection.

Referring to step 108, an EDX spectrum (defect spectrum) is obtained from the defect site. This operation may involve translating the substrate such that the defect site is in the field of view of the SEM; and scanning the primary electron beam over the defect site while detecting the energy spectrum of the x-rays thereby generated. The defect spectrum from each defect site can be stored in one manner to correlate with the location (coordinate) of the defect site.

Referring to step 110, movement from one of the defect sites to one of the reference sites corresponding to the defect site can be performed. As further explained below with respect to FIG. 3, if the defect site is indicated as being in a repeating cell array, movement to the reference site can be achieved by deflecting the primary electron beam to a corresponding position in an adjacent cell. As further explained below with respect to FIG. 5, if the defect site is indicated as being in a non-array patterned structure (ie, the patterned structure is not a repeating cell array), the stage can be translated onto the target substrate by The movement to the reference site is achieved by a corresponding position on one of the adjacent dies. The indication of whether the defect site is in an array or non-array patterned structure can be provided by the data in the resulting file.

Referring to step 112, an EDX spectrum (reference spectrum) is obtained from the reference site. This operation may involve scanning the primary electron beam over the reference site while detecting the energy spectrum of the x-rays thereby generated. The reference spectrum can be stored in a manner to be associated with a corresponding defect site.

Referring to step 114, a difference spectrum can be generated. In one embodiment, normalization of the defect spectrum and a corresponding reference spectrum can be performed, and the normalized reference spectrum can be subtracted from the normalized defect spectrum to obtain a difference spectrum.

Referring to step 116, the element information of the defect can be obtained from the self-difference spectrum. For example, if the selected defect type is a particle type, the element information may indicate the elemental composition of the particle. Have Interestingly, the elemental information obtained using the methods disclosed herein is more accurate than the elemental information obtained from a conventional automated EDX retesting procedure. This is because the method disclosed herein locates the reference site more accurately than the previous method.

An example bar graph is provided to the right of steps 106 through 116. As seen in the example bar graph, the elemental information of the particle defect may indicate that the particle defect has an elemental composition of one of primary (Si) or primary (C) or primary (Fe).

Referring to step 118, more accurate element information obtained using the methods disclosed herein can be combined with the morphological information from step 102. An example bar graph with combined information is provided to the right of step 118. The bar graph shows the defect frequency versus defect type. As shown in the example bar chart, four types of defect bins are depicted. The first two types (#1 and #2) correspond to the types based on the morphology discussed above with respect to step 102 (holes and scratches, respectively). In this case, both the third type (#3) and the fourth type (#4) may be directed to the particles depending on their elemental composition. For example, the fourth type (#4) may correspond to particles of the composition system main (Fe), and the third type (#3) may correspond to particles of other composition (ie, non-Fe). Note that this example is provided for illustrative purposes. Other types of defects are available that have different combinations of morphological information and elemental information.

2 is a flow diagram of one of the program flows 200 for automatically subsampling identified defects in accordance with an embodiment of the present invention. Automatic sub-sampling advantageously reduces the number of defects that EDX re-detection applies to, and thus reduces the time required to perform automatic EDX re-detection. This is because EDX retesting is only applied to a selected type (or several selected types) of defects, not all defects identified in the result file from one of the automated tests.

Referring to step 202, SEM automatic re-detection of the identified defects can be performed. The SEM automatic retest can involve high resolution SEM imaging of defects identified in one of the results files from the automated test.

Referring to step 204, the defect can be classified (i.e., "divided") into a number of defect types. In this example, as shown in the bar chart on the right side of step 204, the defects can be classified into A type of "dust" (particles) defect and a type of "other" (non-particle) defect.

Referring to step 206, one of the defects of the "falling dust" defect type is selected for EDX re-detection. In other words, automatic subsampling selects only dust defects for EDX retest without performing EDX retest on other defects.

Referring to step 208, EDX re-detection is performed on the selected defect in an automated manner. In accordance with an embodiment of the present invention, automatic EDX re-detection can be performed with an accurate reference spectrum using the automated EDX program 105 with accurate reference spectra set forth above with respect to FIG. For example, as shown in the bar graph on the right side of step 208, the dust defects can be classified (divided into) a 矽 (Si) particle type, a bismuth oxide (Si, O) particle type, and a carbon ( C) Particle type.

3 is a flow diagram of one of the program flows for automatic EDX re-detection in an array mode in accordance with an embodiment of the present invention. The array mode can be used to pattern regions on a target substrate having a repeating cell array. For example, the target substrate can include a memory cell array, and the array mode can be used for automatic EDX re-detection within the memory cell array.

Referring to step 302, an automatic check can be performed to detect defects in the target substrate. For example, an automated inspection can be performed using an SEM-based automated inspection device. The inspection tool can be integrated with or separate from the SEM-based automatic re-detection device.

Referring to step 304, the information about the defect detected by the automatic inspection can be stored in a result file. The information contains coordinates about the location of the defect.

Referring to step 306, one of the detected defects in the array can be automatically retested by the SEM. Among other steps, the SEM automatic retest may also involve high resolution SEM imaging of defects identified in the results file from the automated inspection.

Referring to step 308, the defect can be classified (i.e., "divided") into a number of defect types. This classification can be based on the morphology of defects as can be observed from high resolution SEM imaging. For example, the defect type can include a hole type, a scratch type, and a particle ("dust") Types of. These are just a few examples of defect types. Of course, other and different types of defects can be utilized.

Referring to step 310, this step can be performed for each defect in the resulting file. In this step, one of the defects in the result file can be selected as the defect of interest (DOI), and one can make a determination as to whether the EDX re-detection will be applied to the DOI. This determination can be made in an automated manner based on the type of defect of the DOI. If the EDX re-detection is not applied to the DOI, then the processing of the current DOI may end, and the method 300 may loop back and perform step 310 for the next defect (if any) selected as one of the DOIs. On the other hand, if EDX re-detection is to be applied to the DOI, method 300 can proceed to perform steps 312 through 316.

Regarding step 312, EDX is applied to the DOI. This operation may involve translating the stage to move the site of the DOI into the field of view; and obtaining an EDX spectrum from the DOI.

With regard to step 314, EDX is applied to a corresponding location in a reference unit. This operation may involve deflecting the primary electron beam in one direction equal to one of the unit sizes along one of the directions. This deflection causes the primary electron beam to move such that it collides in an adjacent unit (reference unit) corresponding to a position (reference point) of the position of the DOI in its unit. An EDX spectrum can then be obtained from the reference point.

Regarding step 316, a difference spectrum between the EDX spectrum of the DOI and the EDX spectrum of the reference point can be generated. The EDX spectrum can be normalized before the difference spectrum is produced. Thereafter, the difference spectrum can be used to obtain information about the elements of the DOI, as set forth above with respect to step 116 in FIG.

4 shows an electron micrograph of a defect of interest (DOI) 402 and one of the adjacent reference points 404 of one of the adjacent cells, in accordance with an embodiment of the present invention. In this case, the DOI 402 is within a repeating unit array. As such, automatic EDX can be performed in an array mode in which the reference spectrum is obtained by deflecting the electron beam by one of the widths (or heights) of one of the cells.

Figure 5 is a diagram for automatic operation in a non-array mode in accordance with an embodiment of the present invention. One of the flow charts of one of the program flow of EDX retesting. A non-array ("random") mode can be used for regions of a target substrate that are not patterned with a repeating cell array. For example, a special application integrated circuit can include a region with a custom logic circuit, and a non-array mode can be used for automatic EDX re-detection within the region.

Referring to step 502, an automatic check can be performed to detect defects in the target substrate. For example, an automated inspection can be performed using an SEM-based automated inspection device. The inspection tool can be integrated with or separate from the SEM-based automatic re-detection device.

Referring to step 504, the information about the defect detected by the automatic inspection can be stored in a result file. The information contains coordinates about the location of the defect.

Referring to step 506, one of the detected defects can be automatically retested by the SEM. In this case, the defect can cause automatic re-detection in a "random" (non-array) mode in a region that is not a repeating cell array (i.e., a non-array region). Among other steps, the SEM automatic retest may also involve high resolution SEM imaging of defects identified in the results file from the automated inspection.

Referring to step 508, the defect can be classified (i.e., "divided") into a number of defect types. This classification can be based on the morphology of defects as can be observed from high resolution SEM imaging. For example, the defect type can include a hole type, a scratch type, and a particle ("dust") type. These are just a few examples of defect types. Of course, other and different types of defects can be utilized.

Referring to step 510, this step can be performed for each defect in the resulting file. In this step, one of the defects in the result file can be selected as the defect of interest (DOI), and one can make a determination as to whether the EDX re-detection will be applied to the DOI. This determination can be made in an automated manner based on the type of defect of the DOI. If the EDX re-detection is not applied to the DOI, then the processing of the current DOI may end, and the method 500 may loop back and perform step 510 for the next defect (if any) selected as one of the DOIs. On the other hand, if EDX re-detection is to be applied to the DOI, method 500 can proceed to perform steps 512 through 520.

With respect to step 512, electron beam imaging can be performed in a field of view centered on the defect location coordinates of the DOI on the defective die to image a region encompassing the defect site. Preferably, the image surrounding the area of the defect site is a high resolution image.

Regarding step 514, EDX is applied to the DOI to obtain an EDX spectrum (defect spectrum) of one of the DOIs. Elemental information from this EDX spectrum contains not only elemental information from the DOI, but also elemental information from materials surrounding the DOI.

With respect to step 516, electron beam imaging can be performed in a field of view centered at the defect location coordinates on the reference die (rather than the defective die) to image a region encompassing the reference site. Preferably, the image surrounding the region of the reference site is a high resolution image.

After step 514 and prior to step 516, the stage holding the target substrate is translated to move the defect site on the defective die to a reference bit on the reference die (which preferably is adjacent to the defective die) point. In addition, the pattern of images encompassing the reference site can be aligned to the image of the image that encompasses the defect site, and the aligned image can be used to determine the position of the reference site with high accuracy.

Regarding step 518, EDX is applied to the reference site to obtain an EDX spectrum (reference spectrum) from one of the reference points. After normalization, this reference spectrum can be subtracted from the defect spectrum to obtain a difference spectrum, as indicated in step 520. Thereafter, the difference spectrum can be used to obtain information about the elements of the DOI, as set forth above with respect to step 116 in FIG.

6 shows an electron micrograph of a defect of interest (DOI) and a corresponding reference point on a reference die, in accordance with an embodiment of the present invention. In this case, the DOI is in one of the non-array portions of the die. As such, automatic EDX can be performed in a non-array mode in which the reference spectrum is obtained by first locating a reference point on the reference die, and the stage translation is used to move between the DOI and the reference point.

7 is a flow diagram of one of an automated procedure for combining a morphological defect result with an element defect result to obtain a final Pareto, in accordance with an embodiment of the present invention.

Referring to step 702, one of the detected defects can be automatically retested by the SEM. except In addition to the other steps, the SEM automatic retest can involve high resolution SEM imaging of defects identified in one of the results files from the automated test.

Referring to step 704, automatic classification of defects can be performed based on the form of the defect. Defects can be classified (ie, "divided") into a number of defect types, and each defect type can be represented by a category code. For example, the defect type may include one of type 1 as a hole type, one type 2 as a scratch type, and one type 3 ("dust") type. These are just a few examples of defect types. Of course, other and different types of defects can be utilized.

Referring to step 706, a defect of a particular class code can be selected for automatic EDX re-detection while automatic EDX re-detection can be skipped for defects of other class codes. In other words, for each defect, a determination can be made as to whether the defect has a particular class code. If the category code is not a particular category code, then automatic EDX re-detection of the current defect is not performed, and method 700 can loop back and step 706 is performed for a next defect, if any. On the other hand, if the category code is a particular category code, method 700 can proceed to perform steps 708 and 710.

Referring to step 708, automatic EDX re-detection is performed for defects of a particular class code. In order to provide an accurate reference spectrum, if the defect being retested is within a repeating cell array, automatic EDX retesting may be performed by means of steps 312 through 316 set forth above with respect to Figure 3, or if the defect is being retested Not in this array area, automatic EDX re-detection is performed by means of steps 512 to 520. As a result of one of the steps 708, a difference spectrum can be obtained for each defect of the particular class code.

Referring to step 710, the difference spectrum can be used to obtain an elemental composition of each of the defects of the particular class code, and the defects of the particular class code can be binned based on its elemental composition. For example, if a particular category code corresponds to a particle or "dusting" defect, the defects may be separated into a plurality of elemental compositions based on the elemental composition of the defects. For example, the elemental composition may include a 矽 (Si) cell, a bismuth oxide (Si, O) cell, and a carbon (C) cell.

Referring to step 712, a defect Pareto (final Pareto) that is used in classifying the defect and combining the morphological information with the element information can be generated. A bar graph of a defect Pareto system defect frequency pair type. The defect Pareto can be used to make a determination as to what corrective action needs to be taken to reduce the defect rate. In one embodiment, defects having a particular class code can be divided based on the elemental composition of the defect.

For example, consider that the morphological defect type from step 704 includes one of the hole types as Type 1, one of the scratch types of Type 2, and one of Type 3 ("Dust") types. Further consideration is given to the category code (particle or "dusting" defect) of the particular category code system type 3 used to select the defect for the automatic EDX re-detection in step 706. It is further considered that these dust defects can be classified based on the elemental composition of the particles. For example, a dust fall defect of a tantalum oxide can be coded as type 4, while a dust fall defect having other compositions (for example, Si or C) can remain encoded as type 3. The resulting bar graph on the left side of step 712 depicts the final Pareto in this example.

Figure 8 is a diagram of one of the prior art programs 800 for automatic EDX re-detection and one of the inventive programs 850 for automatic EDX re-detection, in accordance with an embodiment of the present invention. As illustrated, the inventive program 850 advantageously requires fewer steps and provides results more quickly than existing programs 800.

The existing procedure 800 includes the steps of: requesting one of the 802 re-detection operations by the operator; transferring 804 the wafer to be re-detected by SEM to the re-detection tool (this can take approximately 15 minutes); performing 806 automatic SEM re-detection (this It can take approximately 10 minutes); the operator performs 808 to divide the defect into offline classifications of several defect types (this can take approximately 30 minutes); the operator classifies a set of defects for a selected type (for example, Dust Defect) Request 810 an EDX job; transfer 812 to the EDX tool by EDX retesting the wafer (this can take approximately 15 minutes); and then perform 814 Auto EDX on the set of defects (this can cost roughly 5 minutes).

As indicated, an operator action is required for the three steps in the existing program 800 (steps 802, 808 and 810). Existing program 850 can typically take approximately 75 minutes until the desired EDX data is obtained.

In contrast, the inventive procedure 850 includes the steps of: requesting an operator to re-detect the work by 802; transferring 804 the wafer to be re-detected by SEM to the re-detection tool (this can take approximately 15 minutes); The SEM retests (this can take approximately 5 minutes); and then the 852 instant EDX retest can be performed using the method 100 (see steps 102 through 118) based on the automatic determination described above with respect to Figure 1 in a general manner.

As indicated, the operator action is required only for the first step in the inventive program 850 (step 802) (rather than the three steps in the existing program 800). Moreover, the inventive program 850 can typically take only about 30 minutes until the desired EDX data is obtained (rather than using the existing program 800 takes about 75 minutes).

9 is a schematic illustration of one of the scanning electron microscope devices 900 having an energy dispersive x-ray (EDX) detection system 920, in accordance with an embodiment of the present invention. As shown in the cross-sectional view of FIG. 9, the electron beam column may include a source 901, a Wien filter, a deflector 906, a focus lens 907, an objective lens 908, and one of the target substrates 910. Taiwan 919.

Source 901 produces an incident electron beam (primary electron beam) 902. The incident electron beam 902 can pass through a one-dimensional filter 904. The weiin filter 904 is configured to generate one of an electric field and a magnetic field that intersect each other. A controllable deflector 906 and a focusing lens 907 can be utilized. The deflector 906 can independently control the electrostatic field in the x-direction and in the y-direction. The deflector 906 can be controlled to cause the electron beam to scan across the surface of the target substrate 910 or to deflect the electron beam for other purposes. For example, the target substrate 910 can be a patterned substrate, such as an integrated circuit fabricated or a refracting refractory mask.

Focusing lens 907 is utilized to focus incident electron beam 902 into a beam spot on the surface of a wafer or other substrate sample 910. According to an embodiment, the focusing lens 907 can be operated by generating an electric field and/or a magnetic field.

As a result of scanning of the incident electron beam 902, secondary electrons are emitted or scattered from the target substrate 910 (which may, for example, be a semiconductor wafer or a scale mask). The target substrate 910 can be held by a movable stage 911.

The secondary electrons can be extracted from the target substrate 910 by the electromagnetic field exposed to the objective lens (final lens) 908. The electromagnetic field is used to trap the emitted electrons to a relatively small distance from the optical axis of the incident beam and to accelerate the electrons upward into the column. In this manner, the primary electron beam 912 is formed from the secondary electrons.

The neutron filter 904 deflects the secondary electron beam 912 from the optical axis of the incident electron beam 902 to a detection axis (the optical axis of the secondary electron (SE) detection system 914 for the device). This is used to separate the scattered electron beam 912 from the incident electron beam 902. The SE detection system 914 detects the secondary electron beam 912 and generates a data signal that can be used to form an image of the surface of the target substrate.

An instrument control and data processing (control/processing) system 950 can include one or more processors (ie, microprocessors or microcontrollers) 952, data storage (eg, including hard disk storage and Memory chip 954, a user interface 957 and a display system 958. The data store 954 can store or maintain computer readable code (instructions) 955 and data 956, and the processor 952 can execute the code 955 and process the data 956. User interface 957 can receive user input. Display system 958 can be configured to display image data and other information to a user.

Control/processing system 950 can be coupled to various components of the electron beam column and can be used to control various components of the electron beam column to implement the methods or programs disclosed herein. For example, movement of stage 911 and scanning by deflector 906 can be controlled by computer readable program code 955 executed by control/processing system 950.

In addition, the control/processing system 950 can also process electronic image data from the SE detection system 914 and x-ray data from the EDX detection system 920. In particular, computer readable code 955 in control/processing system 950 can be used to implement programs related to the automated EDX method as set forth herein.

in conclusion

The drawings set forth above are not necessarily to scale and are intended to be illustrative and not limited to a particular embodiment. The particular size, geometry, and lens current of the magnetic objective will vary and will depend on each embodiment.

For example, the techniques set forth above can be used in an automated inspection and detection analysis system and for inspection and re-detection of wafers, X-ray masks, and the like in a production environment. Other uses are also possible.

In the above description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the invention. However, the above-described embodiments of the present invention are not intended to be exhaustive or to limit the invention to the precise forms disclosed. Those skilled in the art will recognize that the invention can be practiced without one or more of the specific details or by other methods, components, and the like. In other instances, well-known structures or operations are not shown or described in detail to avoid obscuring aspects of the invention. While the invention has been described with respect to the specific embodiments and examples of the present invention, it will be understood by those skilled in the art that various equivalent modifications can be made within the scope of the invention.

These modifications can be made to the invention in light of the above detailed description. The terms used in the following claims should not be construed as limiting the invention to the specific embodiments disclosed in the specification and claims. Instead, the scope of the invention will be determined by the following claims, which are to be understood in accordance with the principles of the claims.

Claims (16)

A method for automatically redetecting a defect detected in a defective die on a target substrate, the method comprising: obtaining a result file of one of the locations including the defects; using a primary electron microscope (SEM) Performing an automatic re-detection of one of the defects to obtain an electron beam image of the defects; performing automatic classification based on the form of the defects determined from the electron beam images to classify the defects into one of several types Selecting a particular type of defect for automatic energy dispersive x-ray (EDX) re-detection; performing the automatic EDX re-detection for the particular type of such defect; and generating one of the distributions of the defect frequency pair type Defect Pareto, wherein the defect Pareto combines form information and element information, and the combination divides certain defects of the specific type into a plurality of defects based on the form of the defects, depending on the element information Types of. The method of claim 1, wherein performing the automatic EDX re-detection comprises, for each defect of the particular type: moving to a defect site; obtaining an EDX spectrum from the defect site; moving from the defect site to a defect a reference site; obtaining the EDX spectrum from the reference site; and generating a difference spectrum from the EDX spectrum from the defect site and the EDX spectrum from the reference site. The method of claim 2, wherein the defect site is indicated as being in a repeating cell array, and wherein the defect site is performed to the reference bit by: The movement of the point: a primary electron beam is deflected by a unit size in one direction such that the reference position is in an adjacent unit corresponding to a position of the defect site. The method of claim 2, wherein the defective site is indicated as being in a non-array patterned structure, the method further comprising: performing electron beam imaging to obtain an area including the defect site on the defective crystal grain a first image; performing electron beam imaging to obtain a second image of one of the regions of the reference site on an adjacent die; and determining the first image after aligning the second image to the first image The position of one of the reference sites in the two images. The method of claim 4, wherein the moving from the defect site to the reference site is performed by: panning and holding one of the target substrates to make a view of the SEM from the defect The defect site on the die moves to the reference site on the adjacent die. The method of claim 2, further comprising: deriving element information from the difference spectrum for each defect of the particular type. The method of claim 1, wherein the defect Pareto comprises a bar graph of a defect frequency pair type. The method of claim 1, wherein the defect Pareto distinguishes between a plurality of holes, a plurality of scratches and a plurality of particles and further distinguishes between a plurality of dust particles of a first composition and a plurality of dust particles of a second composition. A device for automatically redetecting a defect detected on a target substrate, the device comprising: an electron beam column for generating a primary electron beam and focusing the primary electron beam to a surface of the target substrate on; a movable stage for holding the target substrate under the primary electron beam; a deflector for deflecting the primary electron beam; and an electronic detector for detecting a collision due to the primary electron beam And a secondary electron emitted from the surface of the target substrate; an x-ray detector configured to detect x-rays emitted from the surface of the target substrate due to collision of the primary electron beam; and a control system comprising a non-transitory data store for storing computer readable code and data, and further comprising a processor for executing the computer readable code, wherein the computer readable code comprises An instruction to obtain a result file containing a location of the defects detected in one of the defective dies on the target substrate; performing an automatic secondary electron microscope (SEM) Re-detecting to obtain an electron beam image of the defects; performing automatic classification based on the form of the defects determined from the electron beam images to classify the defects into one of several types; a particular type of defect for automatic energy dispersive x-ray (EDX) re-detection; performing the automatic EDX re-detection for the particular type of such defect; and generating a defect that is one of the types of defect frequency pairs The entanglement, wherein the defect Pareto combines morphological information and elemental information, the composition is divided into a plurality of types based on the form of the defects based on the form of the defects. The apparatus of claim 9, wherein the instructions to perform the automatic EDX redetection further comprise instructions for performing the following operations: Moving to a defect site; obtaining an EDX spectrum from the defect site; moving from the defect site to a reference site; obtaining the EDX spectrum from the reference site; and the EDX spectrum from the defect site And the EDX spectrum from the reference site produces a difference spectrum. The device of claim 10, wherein the defective site is indicated as being in a repeating cell array, and wherein the instructions for moving from the defective site to the reference site comprise using a primary electron beam An instruction to deflect a unit size in one direction. The apparatus of claim 10, wherein the defective location is indicated as being in a non-array patterned structure, and wherein the instructions to perform the automatic EDX redetection further comprise instructions to: use an electronic Beam imaging to obtain a first image of one of the regions of the defect site on the defective die; using electron beam imaging to obtain a second image of one of the regions of the reference site on an adjacent die And determining a position of the reference position in the second image after the second image is aligned to the first image. The apparatus of claim 12, wherein the instructions for moving from the defect site to the reference site comprise the stage for translating and holding the target substrate such that a view of the SEM is from the defective crystal The defect site on the particle moves to a number of instructions of the reference site on the adjacent die. The apparatus of claim 10, wherein the instructions to perform the automatic EDX re-detection further comprise instructions for: deriving element information from the difference spectrum for each defect of the particular type. The apparatus of claim 9, wherein the defect Pareto comprises a bar graph of a type of defect frequency pair. The device of claim 9, wherein the defect Pareto distinguishes between a plurality of holes, a plurality of scratches and a plurality of particles and further distinguishes between a plurality of dust particles of a first composition and a plurality of dust particles of a second composition.